Therapy refractory lesions are the main cause of death in solid tumor patients. Although the tumor cells in such lesions express various tumor associated antigens (TAA), they are usually "concealed" from the immune system because of immunosuppressive microenvironment within the tumor, activity of regulatory T (Treg) cells and lack of appropriate presentation of TAA by antigen presenting cells (APC). We propose a novel immunotherapy to destroy such lesions and convert them into endogenous tumor vaccines by intratumoral injection of micelles of glycosphingolipids bearing a-gal epitopes (Gala1-3Galb1-4GlcNAc-R) (designated GSL?gal). This injection induces a strong intratumoral inflammatory reaction due to binding of the natural anti- Gal antibody (1% of immunoglobulins in humans) to a-gal epitopes of GSL?gal. Anti-Gal binding to a-gal epitopes naturally expressed on cells of non-primate mammals (e.g. pigs) induces rapid rejection of xenografts in humans and monkeys. We hypothesize that intratumoral injection of GSL?gal will result in destruction of lesions in a mechanism similar to xenograft rejection and will induce a protective immune response against micrometastases, potent enough to overcome the suppressive effect of Treg cells. The effects of GSL?gal will be mediated by the following mechanisms: 1. Chemotactic factors generation due to complement activation by Anti-Gal/GSL?gal interaction. 2. Insertion of GSL?gal into cell membranes, resulting in anti-Gal mediated destruction of tumor cells as in xenograft rejection. 3. Opsonization of tumor cell by anti-Gal, targeting them for effective uptake by APC recruited into the tumor by the inflammation, thereby converting tumors into endogenous vaccines. Our preliminary studies in knockout mice lacking a-gal epitopes and producing anti-Gal, demonstrated destruction of treated melanoma lesions and their conversion into vaccine in a large proportion of mice. Our specific aims are: 1. To study the mechanisms by which GSL?gal induces inflammation within of B16 melanoma lesions. 2. To evaluate the role of Treg cells in preventing induction of a local and systemic protective anti-tumor immune response in the ~30% of mice that do not respond to GSL?gal treatment. 3. To determine in a model of spontaneous mammary carcinoma the protective effect of intratumoral GSL?gal injection in the development of the primary lesion and on induction of a protective immune response preventing additional lesions and metastases. Success in these studies will enable us to establish clinical trials with GSL?gal in cancer patients with chemotherapy refractory lesions. Our objective is to develop a novel method for treating cancer patients with recurrent chemotherapy refractory solid tumors, by injection into the lesion a substance called glycosphingolipids containing a carbohydrate structure called the a-gal epitope. This substance is referred to as GSL?gal. The treatment exploits the fact that all humans have a natural antibody called anti-Gal in very large amounts. This antibody interacts with a-gal epitopes on GSL?gal and induces local inflammation. Injection of GSL?gal into tumor lesions destroys the lesions with the assistance of anti-Gal and convert treated lesions into a self vaccine. This vaccine "educates" the immune system to recognize tumor cells in micrometastases (small groups of tumor cells in various organs) as cells that have to be destroyed. We will study this treatment in a unique mouse model that simulates the pertinent human immunological parameters. Success in the studies in the mouse model will enable us to apply this treatment to cancer patients.